This paper investigates the resource allocation for two-way relay networks with symmetric data rates from an information theoretic perspective, where a round of information exchange between two sources requiring equal end-to-end transmission rates is considered to be completed by a muti-access (MAC) phase and a broadcast (BC) phase. Decode-and forward (DF) protocol is employed. In this case, we formulate an optimization problem to maximize the sum rate of the system under total available energy. Our goal is to seek the jointly optimized time assignment between the MAC and BC phases and the power allocation among the source and relay nodes. Since the problem is difficult to solve in general, we firstly discuss it in two extreme cases by considering very low and very high system available energy. By doing so, we find an interesting result that in the very high energy case, the optimal ratio of the time assigned for the MAC phase to that assigned for the BC phase is a constant, i.e., 2 : 1. Further, we adopt such constant time assignment to general cases, and derive a closed-form power allocation for two-way relay transmissions. Extensive numerical results vitiated the proposed joint resource allocation and show that the maximum system sum-rate can be approached by our scheme, which obviously excels traditional equal time assignment and equal power distribution schemes.